torch.jit.ignore

torch.jit.ignore(drop=False, **kwargs)

This decorator indicates to the compiler that a function or method should be ignored and left as a Python function. This allows you to leave code in your model that is not yet TorchScript compatible. If called from TorchScript, ignored functions will dispatch the call to the Python interpreter. Models with ignored functions cannot be exported; use @torch.jit.unused instead.

Example (using @torch.jit.ignore on a method):

import torch
import torch.nn as nn


class MyModule(nn.Module):
    @torch.jit.ignore
    def debugger(self, x):
        import pdb

        pdb.set_trace()

    def forward(self, x):
        x += 10
        # The compiler would normally try to compile `debugger`,
        # but since it is `@ignore`d, it will be left as a call
        # to Python
        self.debugger(x)
        return x


m = torch.jit.script(MyModule())

# Error! The call `debugger` cannot be saved since it calls into Python
m.save("m.pt")

Example (using @torch.jit.ignore(drop=True) on a method):

import torch
import torch.nn as nn

class MyModule(nn.Module):
    @torch.jit.ignore(drop=True)
    def training_method(self, x):
        import pdb
        pdb.set_trace()

    def forward(self, x):
        if self.training:
            self.training_method(x)
        return x

m = torch.jit.script(MyModule())

# This is OK since `training_method` is not saved, the call is replaced
# with a `raise`.
m.save("m.pt")